Non-medical drivers of health and pediatric no-show prediction: identifying modifiable risk factors
Related Experiment Video
Updated: Oct 10, 2026

Assessment and Evaluation of the High Risk Neonate: The NICU Network Neurobehavioral Scale
Published on: August 25, 2014
Ashley Gibson1, Linh Nguyen2, Lishan Yu3
1McGovern Medical School, The University of Texas Health Science Center at Houston, Houston, TX, USA. Ashley.R.Gibson@uth.tmc.edu.
Background:
No-shows to medical appointments contribute to disruptions and inefficiencies in healthcare delivery. While artificial intelligence/machine learning (AI/ML) models are increasingly used to predict no-shows, few have incorporated non-medical drivers of health (NMDOH), particularly in pediatric populations. This study evaluated the inclusion of NMDOH in pediatric AI/ML no-show prediction and identified modifiable factors associated with no-shows.
Methods:
A retrospective study was conducted of scheduled pediatric appointments across six general pediatrics clinics during April-August 2024 using electronic health record and NMDOH screening data. A transformer-based AI/ML no-show model was evaluated using area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC). Multivariable logistic regression was used to evaluate factors associated with no-show status.
Results:
Among 7,931 included appointments, 2761 (34.8%) were no-shows. Models with and without NMDOH variables showed similar performance (AUROC 0.708 for both; AUPRC 0.536 vs. 0.535, respectively). Inclusion of patient demographic and insurance variables produced a small improvement in performance, with similar results with and without NMDOH inclusion (AUROC 0.721 and AUPRC 0.542 for both). No-shows were associated with unconfirmed appointments, non-active patient portal status, transportation needs, and housing instability.
Conclusions:
NMDOH variables did not significantly change model performance but identified actionable factors associated with no-shows. Prediction models may be most useful when paired with targeted outreach and barrier-reduction strategies.
Impact:
Non-medical drivers of health (NMDOH), including transportation needs and housing instability, as well as appointment confirmation and patient portal activation, were associated with missed pediatric appointments. Inclusion of NMDOH in AI/ML no-show prediction models did not meaningfully change prediction performance. The associations highlight modifiable barriers that can be targeted through clinical and operational interventions. AI-based no-show prediction may be most effective when paired with outreach and barrier-reduction strategies.
Related Concept Videos
Drug Toxicity: Risk factors
Factors Affecting Illness
For instance, risk factors are connected to illness, disability,...
Attention-Deficit/Hyperactivity Disorder
Diagnostic Criteria and Symptoms
To diagnose ADHD, symptoms must manifest before age 12 and be evident across multiple settings.